A Novel interval type-2 fuzzy algorithm in three-dimensional image reconstruction

Abstract

A crucial part of 3-D image reconstruction is the 3-D image registration method. In this thesis, the focus is on estimating the transformation matrix in geometric alignment strategies. The improvement of optimization algorithms is emphasized to find the best transformation parameters to explain the relationship between two sets of 3-D data. In addition, the proposed algorithms are applied to the uncertain data for 3-D image registration. The experimental data is divided into two types: real number data and fuzzy number data. Therefore, this thesis also proposes the use of interval type-2 fuzzy set based on the optimization algorithms to solve for uncertainty problems. The proposed algorithms are evaluated the performances with benchmark functions, 3-D synthetic datasets, and 3-D real-world datasets.